grimalkina

grimalkina, 8 days ago

Don't vary too many things at once -- think about creating "resilient" interactions with a generative tool by including elements you know that YOU understand well or already know the answer to along with the unknown, and think about how you're building a way for yourself to test a translation, implementation, or new method. These tools are statistical so think about what a good question for a statistical model is (what is MOST common, MOST prevalent, MOST expected by the way this is trained)

grimalkina, 8 days ago

Overall compare and contrast is incredibly good for our learning skills but we tend to avoid it! We often think "well if I want to get good at a thing I should use it all the time or ONLY focus on it." Not true. Compare and contrast unlocks many superpowers in learning and is really efficient. Give yourself permission to HALF use a tool like copilot and half not.

grimalkina, 8 days ago

Let's move away from individual cognition and into the social (SPOILER, they interact!!!). Our research has found that a LARGE % of developers don't KNOW what tools their colleagues use and report mixed usage. This creates ambiguity, confusion, and my guess is often a lot of fear esp for junior folks wondering if they "look like a good developer in this org."

Simply raising the question: "what are our biggest fears about genAI?" can be POWERFUL

grimalkina, 8 days ago

Doing this as a group conversation is a hallmark of a learning and belonging culture (https://www.pluralsight.com/resource-center/guides/new-developer-research-paper). Our research also found that you shouldn't assume "everyone is doing it", team vs individual usage is mixed. Who is "left out" of adoption is an important conversation here. Shared team norms is probably a really good place to start -- answering the q "what do I expect from you as my colleague with these tools" is a meaningful conversation

grimalkina, 8 days ago

Whether or not we SHOULD use these tools, the fact that large numbers of developers ARE using these tools suggests to me that they probably need help understanding common misconceptions that we solidly know people make when it comes to efficient or inefficient problem-solving strategies

Rarely has this stuff been translated into software but it's increasingly a theme for me this year to return to "learning," my first love, the most disregarded, and undervalued, and needed science

grimalkina, 8 days ago

It's really odd that generative AI has captured the world on this, because many of the questions are questions we have been asking and could imagine different futures for. For example I was reading about and experimenting (very unsuccessfully) with personalized tutors in classrooms when I was a postdoc. Absolutely old idea, raising MANY of the same questions as we raise now. What about error? In what ways would a personalized tutor benefit a student? But would it create a system where some...

grimalkina, 8 days ago

...have "real" support and other students get shunted into "automated" support because that's what's easy? THESE conversations happened as far back as my advisors' advisors. Not saying we aren't grappling with new technology but in many ways I see genAI has become the ONLY example of these larger questions with little awareness that these have been questions for a long time about human work, human learning, and what helps humans.

grimalkina, 12 days ago

@rdnielsen at any rate I was making a point about within-field ability beliefs more than between field; even if not all but SOME ecologists have very sharp stats training and expertise the point still stands because it's about what we exclude from the realm of possibility because of an attribution about a certain area. Across all scientific fields the move away from NHST and into better applied methods has been difficult, but health is not winning on this one

grimalkina, 12 days ago

I've been so interested in hierarchies of knowledge lately and what supposedly observed (not really) latent abilities are assumed to drive the activities that we see. Interesting intersectional cases of this are amazing especially when you can see the judgment flip. I see it daily for my wife depending on whether people are activating their "neuroscientist" stereotypes or "teacher" stereotypes

grimalkina, 12 days ago (edited 10 days ago)

Ecologists do some of the sharpest statistics around because they've had to solve for observational sampling and strength of evidence in the real world against massive constraints but people always think if you worked at a chemistry bench you're somehow a hardcore scientist with objective purity of thought but if you've done field work you just must "care so hard" about the natural world

grimalkina, 8 days ago

@JeffGrigg amazing!!

grimalkina, 8 days ago

This message brought to you by the fact that I just sewed little gold chains into a blazer..... Blazer jewelry isn't really a thing but it should be and no one can stop you

grimalkina, 9 days ago

If your manager doesn't have a good understanding of what learning a new technical skill looks like, will your manager recognize when it's happening?

grimalkina, 11 days ago

@kf perfect 🥹🥹🥹🥹

grimalkina, 10 days ago

The human body and all its systems is so massively complicated and we can measure some things happening in it way better than others and then there are a million billion constraints and proxies we have to rely on and obviously you have to do it all while actively working to do no harm. All respect to people innovating in how we even test in this world because it must be so thankless and invisible to most of us.

grimalkina, 11 days ago

Also happy Pride month...!

Because it is a far cry from my first gig in tech where someone told me to never come out in this industry to now be a proudly out queer leader who incorporates the important perspective of LGBTQ+ developers (a pov so rarely explored in this research) in findings, and gets to present such work to a big highly technical audience. The world is tough but the world is trying. Grateful for expansive communities that welcome me & redefine who is allowed to be "technical"

grimalkina, 11 days ago

@danilo ugh "how do I fix burnout" posts from people making me so so sad as I think about this

grimalkina, 11 days ago

@danilo although I DO believe in the power of small protective communities/healing before "systems" for people. But yeah I'm reading this infrastructure book right now making the point that we are all networked with everyone on the planet and I'm like, "psychological wellbeing is part of this! 😭 "

grimalkina, 11 days ago

@johana I think that is really profound. Naming things IS some sort of solution (not to the world but to ourselves??) if I am following correctly...I think about how it keeps us from participating in our own exclusion because we refuse the bad faith interpretation of what is happening around us and causing us suffering?

grimalkina, 10 days ago

@pencilears I think there's a lot of that for sure